DEGRADATION OF DEXTRANS BY ENZYMES OF INTESTINAL BACTERIA'

THEODORE W. SERY2 AND EDWARD J. HEHRE

Department of Microbiology and Immunology, Cornell University Medical College, New York, New York

Received for publication August 4, 1955

In a previous report we (Hehre and Sery, 1952)showed that large numbers of dextran-splittinganaerobic bacteria, belonging to the Bacteroidesgenus, occur in the human colon and account forthe dextran-degrading capacity of feces firstnoted by Engstrand and Aberg (1950). Examina-tion of a series of isolates of these intestinal ana-erobes revealed that, although growing culturesutilize dextran with the production of acid, sterilefiltrates from dextran-broth cultures depolymer-ize the polysaccharide with the liberation of somereducing sugar (Hehre and Sery, 1952). We havenow examined more closely the soluble dextran-splitting enzyme system produced by a selectedstrain of intestinal bacteroides. This system dif-fers from most previously described dextranases.It hydrolyzes dextrans, apparently in endwisefashion under certain circumstances, to dextroseas the sole or major product.

Dextran-splitting enzymes elaborated by sev-eral types of fungi (Nordstrom and Hultin, 1948;Huiltin and Nordstrom, 1949; Tsuchiya et al.,1952a; Whiteside-Carlson and Carlson, 1952;Kobayashi, 1954), and by the bacterium, Cell-vibrio fulva (Ingelman, 1948), have been foundto cause rapid lowering of viscosity and molecu-lar weight of the substrate before releasing muchreducing sugar. The ultimate products of thistype of degradation, as shown by Jeanes and herassociates (1952, 1953) for dextranase from Peni-cillium funiculosum strain 1768, are oligosac-charides (isomaltose, isomaltotriose, etc.) ratherthan dextrose. Tsuchiya et al. (1952a) have re-ported dextran hydrolysis to dextrose by anamylase preparation from Aspergillus niger strain330, and have otherwise contrasted the activitiesof the A. niger and P. funiculosum systems to-ward several substrates, including three dextransof different chemical composition. In the presentwork, the bacteroides system has been examined

1 Support was received from the Corn IndustriesResearch Foundation.

2 Present address: Research Department, WillsEye Hospital, Philadelphia, Pennsylvania.

in some detail for its action on a number of struc-turally different dextrans and related saccharides.The importance of enzymatic studies in the de-velopment of knowledge of the starch-glycogenpolysaccharides suggests that such studies may,similarly, provide new information on the closelyrelated dextran polysaccharides.

MATERIALS

Bacteroides enzyme preparation. From 24 strainsof enteric dextran-splitting bacteroides, one (FA-1A) was selected for study because it yieldedsomewhat more potent enzyme preparations thanthe others. Stock cultures were maintained on"chocolate" blood agar, i. e., on a meat infusion,yeast extract agar to which 5 per cent sterilecitrated human blood was added, and the wholeheated at about 80 C until brown.

Dextran-splitting enzymes were prepared fromFA-1A cultures in broth comprising: dextran,32 per cent; tryptose, 1 per cent; yeast extract,0.2 per cent; NaCl, 0.5 per cent; KH2PO4, 0.25per cent; and Na2HPO4, 0.68 per cent. Suitableanaerobic conditions for growth were obtainedas follows: to 800 ml of medium, autoclaved in a1-L flask and cooled to about 70 C, sufficientseparately sterilized medium was added (whilehot) to partly fill the neck of the flask.4 The me-dium was then cooled to 37 C and inoculated,

3 The dextran in the medium, essential to in-duction of the enzyme, was the high molecularweight polymer (512-E) formed from sucrose bycell-free dextransucrase of Leuconostoc mesen-teroides NRRL B-512 (Tsuchiya et al., 1952b); itpossessed 95 per cent "terminal plus 1,6-linked"units, as calculated from behavior on periodateoxidation.

4This procedure can be simplified by the use ofa special culture vessel made by sealing the mouthof a 1-L Florence flask into an opening in the baseof a second similar flask. Medium is added to fillthe lower reservoir, with a slight excess (about 50ml) extending into the upper reservoir; the systemmay then be autoclaved without danger of over-flow.

373

on June 8, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

SERYIAND HEHRE

3.0 5.0 6.05.4

4.8

-Ji

>-2.0 /s~~~~~~~~~76

aI.

20 40 f0 80 100TIME, MINUTES

Figure 1. Increase in specific fluidity (reciprocalof specific viscosity) of enzyme-dextran mixturesincubated at different pH.

PR

Figure 2. pH-activity curves for the Bac-teroides enzyme system, with activity representedas rate of increase in specific fluidity *, or as re-

ducing sugar liberated (0, t) during 5 or 8 hoursincubation at 30 C.

using fresh cultures grown 2 or 3 days in an

anaerobic jar; the bacteria from one or two "choc-olate" blood agar plates were suspended in a

small amount of the 2 per cent dextran broth andintroduced into the main flask without mixing;less viscous suspension fluids permitted the in-oculum to rise to the surface of the culturemedium.

After incubation at 37 C for 18 to 24 hr, eachL of culture was treated with 375 g of ammoniumsulfate, held at 4 C for 1 hr, then centrifuged inthe cold. The sediment was taken up in 50 ml ofdistilled water, the bacterial cells separated bycentrifugation, and the clear fluid stored at minus.70 C. Generally, 1 ml of this enzyme solution,when incubated with an equal volume of 1 percent dextran 512-E for 2 hours at 30 C and pH7.4, caused the release of 1.5 to 2.0 mg of reducing,sugar (as dextrose).

EXPERIMENTAL RESULTS

Divergent effect of pHI on enzyme activity. Theaction of the bacteroides enzyme system on dex-tran was early found to be complex in the sensethat rapid reduction of viscosity is the mostprominent effect produced in systems held atlow pH, whereas reducing sugar liberation is thechief feature in systems incubated at higher pH.The influence of pH on the capacity of the

bacteroides system to lower dextran viscosity wasexamined with a series of mixtures comprisingequal volumes of enzyme (1:4 dilution) and of1 per cent dextran 512-E in different buffer solu-tions (acetate for the range pH 3.8 to 5.4, phos-phate for the range pH 5.6 to 7.8). Each mixturewas introduced into a No. 100 Ostwald-Cannon-Fenske viscometer held at 30 C, and measure-ments of the outflow times made during a 2-hrperiod. Figure 1 shows the family of straight linesobtained by plotting, for each mixture, the spe-cific fluidity, sp sp (= 1/X1 sp), against reactiontime (see Hultin and Nordstrom, 1949; Ingelman,1948). Since the individual lines tend to a com-mon origin, their slopes were used as indices ofrelative enzyme activity (Levinson and Reese,1950) in constructing a pH-activity curve (fig-ure 2). It is evident that viscosity-depressingactivity is maximal in the region of pH 5.4, dropsoff sharply at lower pH values, and reaches zeroin the region of pH 4.4; the drop above pH 5.4 ismore gradual, but the enzyme shows only slightactivity at pH 7.6.The effect of pH on ieducing sugair liberation

is quite different. Mixtures of enzyme and dex-tran 512-E, prepared as above but with enzymediluted 1:2, were examined after 5 and 8 hoursat 30 C. Dextrose was the only sugar detected onchromatograms prepared on Wthatman no. 1 filterpaper, using multiple ascents in 6:4:3 n-butanol:pyridine:water (Jeanes et al., 1951) and a sprayof ammoniacal silver nitrate (Hough, 1950). From

374 [VOL. 71

on June 8, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

1]EXTRAN DEGRADATION BY INTESTINAL BACTERIA

the quantities of dextrose (Hagedorn and Jensen,1923) formed in the mixtures at different pHvalues, pH-activity curves of a second type were

constructed. As figure 2 illustrates, the release ofreducing sugars was maximal in the region ofpH 7, where the viscosity-depressing effect of theenzyme was poor; and poor in the range of pH5.0 to 5.5, where the viscosity-depressing effectwas maximal.The divergent influence of pH on the two ac-

tivities was confirmed by tests on a number ofdifferent native and clinical dextrans (table 1).Paired mixtures containing equal volumes of en-

zyme (1 :2 dilution) and of 1 per cent dextran in0.1 M phosphate buffer, pH 5.4 or 7.4, were ex-

amined for changes in viscosity during the firsthr, and for reducing sugar liberation after 6 hrat 30 C. The rate of increase in fluidity, in thoseinstances where it could be measured, was from2 to 4 times higher at pH 5.4 than at pH 7.4,whereas the amount of reducing sugar releasedwas greater at pH 7.4 than at pH 5.4 for all butone of the 10 dextrans tested. This biphasic pH-activity relationship is hardly explainable exceptby the presence of two distinct dextran-splittingenzymes in the bacteroides preparation.

Extent of degradation of structurally differentdextrans. The release of dextrose from differentdextrans, under conditions (pH 7.4) providingrelatively slight initial loss of viscosity, stronglysuggests a preferential enzymatic attack on theouter parts of the dextran molecules. If so, theextent of hydrolysis of a particular dextran couldbe greatly influenced by such structural featuresas the number and length of its outer chains,extent and type of branching, etc. The availabil-ity of a number of purified dextrans differing fromeach other in behavior toward periodate oxida-tion, and in other respects, enabled this point tobe examined. In the following experiment, thecourse and extent of hydrolysis of eight struc-turally different dextrans was followed by meas-

urements of the dextrose liberated and of theresidual polysaccharide.The test mixtures comprised: enzyme, 4 ml;

2 per cent dextran, 2 ml; and 0.2 M phosphatebuffer (pH 7.4) containing 400 units of penicillinper ml, 2 ml. (Pretesting showed that penicillindoes not affect enzyme activity.) Control mix-tures of each dextran without enzyme, and of en-

zyme plus buffer, were incubated with the testmixtures at 30 C. Analyses for reducing sugar re-

lease were made at 6, 24, and 48 hr; determina-

TABLE 1Influence ofpH on the enzymatic capacity to increase

the fluidity of and to retease dextrose fromdifferent dextrans

Rate of Increase Rate of Releaseof Specific of Reducing

Dextran* Fluidityt Sugart

pH 5.4 pH 7.4 pH 5.4 pH 7.4

Native 1 28 7 6 162 23 5 6 113 6 3 4 144 6 2 3 25 8 136 1 57 1 2

Clinical 8 9 199 9 17.10 9 11

* Dextrans 1, 4, 6, 7, kindly supplied by Dr.Allene Jeanes, were from Leuconostoc mesen-teroides NRRL B-512, -1298, -742 and -1355;dextrans 2, 3, 5 were from L. mesenteroides strainB (Sugg and Hehre, 1942), Acetobacter capsulatumNCTC 4943 (Hehre, 1951), and Streptococcus sp.strain 50 (Hehre, 1952): dextrans 8, 9, 10 wereproducts of Commercial Solvents Corporation(Terre Haute, Indiana), A. B. Pharmacia(Uppsala, Sweden), and Dextran Corporation(Yonkers, New York).

t X 103; determined only for mixtures con-

taining dextrans of high initial viscosity.$ Mg reducing sugar liberated in 6 hr per 10 ml

enzyme-dextran mixture; dextrose was the onlysugar detected on chromatograms of the mixtures.

tions of polysaccharide precipitable by 40 percent (by volume) methanol, and of that pre-cipitable in the range of 40 and 65 per centmethanol, were made at 48 hrs. The methanol-treated mixtures were kept at 4 C for 2 hr, thencentrifuged in the cold; the sediments weredrained, dried, and hydrolyzed in sealed tubes byN HCl at 120 C for 20 min. From the dextrose con-tents (Hagedorn and Jensen, 1923) of the neu-tralized hydrolyzates, the percentage of each ofthe polysaccharide residues was calculated, takingas 100 per cent the polysaccharide content of thecorresponding dextran-buffer control. The resultsof the tests at 48 hr are summarized in table 2.

It is evident that the different dextrans weredegraded to different degrees and, more impor-tant, that a general correlation exists between thedegree of degradation on prolonged incubation

1956] 375

on June 8, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

SERY AND HEHRE

TABLE 2Degradation of structurally different dextrans by the bacteroides enzyme system at pH 7.4

Dextran* Residual PolysaccharideDextrose

Terminal + Releasedt Precipitated PrecipitatedSource 1, 6-linked by 40-65 per by 40 per

dextrose units cent methanol cent methanol

per cent per cent per cent per cent

Leuconostoc mesenteroides, NRRL B-512-E........ 95 100 0 0Acetobacter capsulatum, NCTC 4843 (Hehre, 1951).. 90 104 0 0L. mesenteroides, B (Sugg and Hehre, 1942) ......... 91 100 2 0L. mesenteroides, NRRL B-1382 .................. 81 77 2 21L. mesenteroides, NRRL B-1196 .................. 80 56 0 40L. mesenteroides, NRRL B-742 ................... 73 47 4 49L. mesenteroides, C (NRRL B-1298)j.............. 61 42 40 13L. mesenteroides, NRRL B-1355 ........ .......... 58 12 0 90

* Dextran 512-E was the enzymatically synthesized sample, kindly furnished by Dr. Henry M.Tsuchiya, used to induce the bacteroides enzyme. Preparations 1382, 1196, 742, C (or 1298) and 1355were kindly supplied by Dr. Allene Jeanes, who also provided the periodate oxidation data from whichthe percentages of terminal plus 1,6-linked glucose units were calculated.

t Calculated from the amount of reducing sugar liberated through enzymic action, compared withthe amount of released on acid hydrolysis of the corresponding dextran control. Chromatograms of allmixtures showed a heavy spot migrating as dextrose (ca 30 cm); in enzyme mixtures containing dextransB and C, a trace spot (RG = 0.2) of oligosaccharide other than isomaltose was also observed.

t Dextrans of strain C (Neill et al., 1941; Hehre, 1943) and NRRL-B-1382 are characterized by strongreactivity with type 12 pneuimococcus antiserum.

with enzyme, and the percentage of "terminalplus 1,6-linked" glucose units in the molecule asdetermined from periodate oxidation. Both the512-E dextran and the dextran from A. capsu-latum strain 4943, with 96 and 90 per cent "ter-minal plus 1,6-linked" units, were completelysplit to dextrose. Likewise, dextran B, comprising91 per cent of such units, was nearly completelyhydrolyzed-traces only of oligo- and polysac-charide resisting enzymic action. The remainingdextrans, containing lower proportions of "ter-minal plus 1,6-linked glucose" units (i. e., 81, 80,73, 61 and 58 per cent), were hydrolyzed to re-spectively lesser degrees (77, 56, 47, 42, and 12per cent). The polysaccharide residues of theseincompletely hydrolyzed dextrans were, with oneexception (dextran C), materials mostly precipi-tated from solution by 40 per cent methanol,rather than by 40 to 65 per cent methanol. Per-sistence of a high molecular weight "resistantcore" might be expected from a preferential en-zymatic attack on the outer portions of highlybranched dextran molecules, with little or noscission of internal linkages.

In the exceptional case of dextran C, the bulkof the polysaccharide residue was precipitated inthe range of 40 to 65 per cent methanol, rather

than by 40 per cent methanol. This particulardextran would appear to have an internal struc-ture quite different from that of the other dex-trans tested. The unusual type of serological ac-tivity of dextran C (Neill et al., 1941; Hehre,1943) does not seem correlated with its anomalousbehavior towards enzymatic action; dextran 1382,with serological properties of the same type as C,yielded a residual polysaccharide that was pre-dominantly precipitable by 40 per cent methanol.The data in table 2 refer to enzyme-dextran

mixtures incubated 48 hr. That cessation of en-zyme action had not invariably occurred in thatperiod, however, is indicated by figure 3, whichshows the course of glucose liberation in the sameenzyme-dextran mixtures. Three of the prepara-tions (512-E, 4943, B) were hydrolyzed rapidlyand essentially completely in 48 hr. However,four others (1382, 1196, 742, C) were hydrolyzedmore slowly and steadily, and did not reach astationary level of degradation in 48 hr. It is un-certain whether these dextrans would eventuallybe hydrolyzed completely or leave truly refrac-tory, "limit polysaccharides."With dextran 1355 one can be more certain

that a limit polysaccharide was produced, sincedextrose liberation ceased entirely after 6 hr (al-

376 [VOL. 71

on June 8, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

DEXTRAN DEGRADATION BY INTESTINAL BACTERIA

though the enzyme was found active after 48 hr).This dextran differs from all the others testedin possessing an unusually high proportion (35per cent) of dextrose units, presumably 1,3,6-linked branch points, resistant to oxidation byperiodate (Jeanes et al., 1954).The complete hydrolysis of dextran from Aceto-

bacter capsulatum strain NCTC 4943 is worthy ofspecial comment since this dextran, unlike all theothers, was synthesized from dextrins rather thanfrom sucrose.

Enzyme action on amylopectins and glycogens.Corn and potato amylopectins, and oyster and

Figure 3. Release of reducing sugar (as dex-trose) from structurally different dextrans (table2) during incubation for 48 hr with the enzyme.

TABLE 3Degradation of amylopectins and glycogens by the

bacteroides enzyme system

ResidualPolysaccharide

Polysaccharide Released* Precipi- Precipi-Rlae*tated by tated by

40-65 40per cent per centmethanol methanol

per cent per cent per cent

Corn amylopectint.... 22 2 79

Potato amylopectin. .. 26 2 74Oyster glycogen.... 6 2 90

Rabbit liver glycogen. 4 2 92

* Calculated from the amount of reducing sugar

liberated through enzymatic action, comparedwith the amount released on acid hydrolysis of thecorresponding polysaccharide control. Paperchromatograms in every instance showed a singleheavy spot migrating as dextrose (ca 30 cm).

t Pentasol-soluble fraction, kindly supplied byDr. T. J. Schoch.

rabbit liver glycogens were tested as substrates,following the procedure of the preceding experi-ment. Table 3 shows that prolonged (48-hr) incu-bation with the enzyme at pH 7.4 caused notablygreater hydrolysis of the amylopectins (22 and 26per cent) than of the glycogens (4 and 6 percent). In all cases, the only sugar detected chro-matographically was dextrose, while nearly allof the residual polysaccharide was precipitableby 40 per cent methanol and, hence, evidentlyof high molecular weight. Supplementary meas-urements, made with the Klett-Summerson pho-tometer and green filter no. 54, showed retentionof iodine-coloring capacity after enzymatic treat-ment (as compared with the capacity beforetreatment) corresponding to 46 per cent for po-tato amylopectin, 56 per cent for corn amylo-pectin, 81 per cent for oyster glycogen, and 86per cent for rabbit liver glycogen.Enzyme action on oligosaccharides. Tests for

susceptibility to enzyme action, under the sameconditions, also were made of several oligosac-charides possessing an unsubstituted (terminal)D-glucopyranosyl component. Paper chromato-grams and reducing sugar measurements, at vari-ous intervals of enzyme-substrate incubation,

TABLE 4Action of bacteroides enzyme on dextrose-

terminated saccharides

Sugars onChromatograms,

Substrate* Extent of 48-hr MixturesSubstrate* Hydrolysist

Sub- Glucosestrate

Isomaltose Complete in 3 0 +hr

Maltose, panose Complete in 48 0 +hr.

Turanose, sucrose None detected + o0cellobiose, tre- in 48 hrhalose, glucose-1-phosphate

* Isomaltose, crystalline panose, and crystal-line turanose were kindly supplied by Dr. AlleneJeanes, Dr. S. C. Pan and Dr. Nelson K. Richt-myer.

t Determined from the amount of reducingsugar released through enzymic action, comparedwith the amount released on acid hydrolysis.

t No increase over the trace of glucose presentin the enzyme solution; no fructose detected inenzyme-turanose or -sucrose mixtures.

3771956]

on June 8, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

SERY ANI) HEHRE

showed that isomaltose, maltose, and panose were

completely hydrolyzed by the enzyme (table 4).Isomaltose, possessing the structural linkage thatpredominates in dextrans, was the most rapidlyhydrolyzed. In the case of panose, chromato-grams showed the presence of maltose and dex-trose (but not isomaltose) during the first 6 hr ofincubation; at 48 hr only dextrose was present.Thus, the a-1 ,6-glucosidic bond of panose, likethat of isomaltose itself, is more rapidly split bythe enzyme system than is the a-1,4-bond ofpanose and maltose. The lack of any significanthydrolysis of turanose, trehalose, sucrose or glu-cose-i-phosphate indicates that the bacteroidesenzyme is not an a-glucosidase of broad speci-ficity.

DISCUSSION

The bacteroides system is distinguished fromnearly all previously described dextranases on thebasis that dextrose is the sole or dominant sugar

product. The enzyme system of Aspergillus nigerstrain 330 (Tsuchiya et al., 1952a) alone resemblesthe bacteroides system in this respect, being simi-lar also in its great affinity for isomaltose as a

substrate. The two differ, however, in severalimportant respects. The maximal release of glu-cose is at pH 7.0 to 7.5 with the bacterial enzyme,but at pH 4.4 with the mold preparation; also,starch (at least its major component, amylo-pectin) is hydrolyzed less rapidly and completelythan most dextrans by the bacteroides system,but much more rapidly than the dextrans by theA. niger amylase (Tsuchiya, personal communica-tion); finally, transglucosylase activity for mal-tose and cellobiose, not detected with the bac-terial preparation, is a prominent feature of theA. niger 330 system (Tsuchiya et al., 1950). Invork yet unpublished, W. L. Bloom and his asso-

ciates (personal communication), have obtainedfrom the duodenal and jejunal mucosa of severalspecies, including man, another dextranase thatyields dextrose as the major hydrolytic product.It will be of great interest to compare the proper-

ties of this intestinal mucosal enzyme with thepresent system from one of the predominanttypes of intestinal bacteria. The bacteroides en-

zyme, otherwise, appears clearly different fromthe "debranching" enzymes and a--1,6-glucosida-ses of various sources, which cleave a-I ,6-gluco-sidic linkages in starch-type polysaccharides or inoligosaccharides but not in dextrans.

Although knowledge of the structures of poly-saccharides of the starch-glycogen class has beengreatly enriched by studies on the enzymic degra-dation of these substances, only one previous re-port has dealt with the relationship betweenenzymic breakdown and the structure of dex-trans. Tsuchiya, Jeanes, Bricker and Wilham(1952a) compared the rates of sugar liberation,by enzymes from P. funiculosum strain 1768 andA. niger strain 330, for three different dextrans.Both enzymes attacked B-512 dextran (95 percent 1,6-linkages) more rapidly than B-742 dex-tran (75 per cent 1, 6-linkages). The enzyme fromP. funiculosum degraded the third dextran, B-523(75 per cent 1, 6-linkages), at the same rate asB-512 dextran,5 while the enzyme from A. nigerhydrolyzed it at a rate intermediate betweenthose of B-512 and -742.With the Bacteroides enzyme, good correlation

was found between the degree of hydrolysis ofdifferent dextrans and their content of "terminalplus 1,6-linked" glucose units determined byperiodate oxidation. Furthermore, those dextranswith lower proportions of units linked in thismanner, after prolonged incubation with theenzyme, yielded residues precipitable by 40 percent methanol and, presumably, of high mo-lecular weight. These features strongly indicatethat the bacteroides enzyme, at pH 7.4, preferen-tially attacks the outer portions of dextranmolecules. They also furnish independent sup-port for the concept, derived from chemicalstudies, that certain dextrans are highly branchedmolecules. Dextran 1355, for example, appearshighly branched not only because of its highcontent (35 per cent) of dextrose units refractoryto oxidation by periodate and low content (58per cent) of "terminal plus 1,6-linked" units(Jeanes et al., 1954), but also because it yieldeda "limit" dextran, resistant to the enzyme, afteronly 12 per cent hydrolysis.The complete or nearly complete breakdown

to dextrose of several dextrans composed mainlybut not exclusively of 1,6-linked units, as wellas the cleavage of dextrose from the long outerchains of corn and potato amylopectins, indi-cates that the action of the bacteroides systemis not confined to hydrolysis of a-1,6-glucosidic

5 )r. Jeanes (personal communication) statesthat P. funiculosurm 1768 dextranase hydrolyzesdextrans having 96 to about 70 per cent 1,6-link-aiges in proportion to this percentage.

378 [VOL. 71

on June 8, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

DEXTRAN D)EGRADATION BY INTESTINAL BACTERIA

linkages. Tests with isomaltose, panose, andmaltose have provided clear evidence that thespecificity extends to the a-1,4-glucosidic link-age, although cleavage of the a--1,6-linkage isdecidedly more rapid.

ACKNOWLEDGMENTS

The authors are especially indebted to Dr.Allene Jeanes of the Northern Utilization Re-search Branch, United States Department ofAgriculture, for samples of isomaltose and ofperiodate-characterized dextrans used as sub-strates in this study; and to Dr. Henry M.Tsuchiya, of the same Laboratory, for a generoussupply of enzymically synthesized dextran usedto induce the bacteroides enzymes.

SUMMARY

A soluble enzyme system, capable of hydrolyz-ing dextrans to dextrose as the sole or majorproduct, has been obtained from an intestinalbacterium of the Bacteroides genus. This systemevidently contains two different dextranases,since it can either "liquify" or "saccharify"dextrans. (At pH 5.0 to 5.5, the principal effectis rapid lowering of viscosity; at pH 7.0 to 7.5,the major effect is rapid release of glucose.)At pH 7.4 the bacteroides system was found

able to hydrolyze both a--1,6-and a-1,4-gluco-sidic linkages in oligosaccharides, and to causethe complete or nearly complete hydrolysis ofdextrans containing 90 to 95 per cent "terminalplus 1,6-linked" glucose units. Dextrans withsmaller proportions of such units (as well asamylopectins and glycogens) were partiallyhydrolyzed to dextrose under the same condi-tions. Since these incompletely hydrolyzed sub-stances, with one exception, yielded polysac-charide residues readily precipitated by alcohol,it would appear that the enzymic attack wasgenerally limited to the outer portions of themolecules. The exceptional dextran (from Leu-conostoc mesenteroides strain C) presumably hasan internal structure quite different from thoseof the other dextrans examined and, hence, isworthy of further chemical study.

REFERENCES

ENGSTRAND, L. AND ABERG, B. 1950 Eixeretionof intravenously administered dextran. Lan-cet, 1, 1071-1073.

HAGEDORN, H. C. AND JENSEN, B. N. 1923 ZurMikrobestimmung des Blutzuckers mittelsFerricyanide. Biochem. Z., 135, 46-58.

HEHRE, E. J. 1943 Serological properties ofproducts synthesized from sucrose by en-zymes from different strains of Leuconostocbacteria. Proc. Soc. Exptl. Biol. Med., 54,18-19.

HEHRE, E. J. 1951 The biological synthesis ofdextran from dextrins. J. Biol. Chem., 192,161-174.

HEHRE, E. J. 1952 An unusual Streptococcusthat yields low molecular weight (clinicaltype) dextran directly. Bacteriol. Proc.,1952, 23.

HEHRE, E. J. AND SERY, T. W. 1952 Dextran-splitting anaerobic bacteria from the humanintestine J. Bacteriol., 63, 424-426.

HOUGH, L. 1950 Application of paper chroma-tography to the separation of polyhydric alco-hols. Nature, 165, 400.

HULTIN, E. AND NORDSTRO5M, L. 1949 Investi-gations on dextranase. I. On the occurrenceand assay of dextranase. Acta Chem. Scand.,3, 1405-1417.

INGELMAN, B. 1948 Enzymatic breakdown ofdextran. Acta Chem. Scand., 2, 803-812.

JEANES, A., HAYNES, W. C., WILHAM, C. A.,RANKIN, J. C., MELVIN, E. H., AUSTIN, M. J.,CLUSKEY, J. E., FISHER, B. E., TSUCHIYA, H.M., AND RIST, C. E. 1954 Characterizationand classification of dextrans from 96 strainsof bacteria. J. Am. Chem. Soc., 76, 5041-5052.

JEANES, A., WILHAM, C. A., JONES, R. W.,TSUCHIYA, H. M., AND RIST, C. E. 1953 Iso-maltose and isomaltotriose from enzymic hy-drolyzates of dextran. J. Am. Chem. Soc.,75, 5911-5915.

JEANES, A., WILHAM, C. A., JONES, R. W., ANDRIST, C. E. 1952 Isomaltose and isomalto-triose from enzyme hydrolyzates of dextran.Abstracts of papers, 121st National A. C. S.meeting.

JEANES, A., WISE, C. S., AND DIMLER, R. J. 1951Improved techniques in paper chromatogra-phy of carbohydrates. Anal. Chem., 23, 415-420.

KOBAYASHI, T. 1954 Dextran. IV. Dextran-destroying enzyme of molds. J. Agr. Chem.Soc. Japan, 28, 352-357. (Chem. Abs., 48,12194g.)

LEVINSON, H. S. AND REESE, E. T. 1950 En-zymatic hydrolysis of soluble cellulose deriva-tives as measured by changes in viscosity. J.Gen. Physiol., 33, 601-628.

NEILL, J. WI., SUGG, J. Y., HEHRE, E. J., AND

3791956]

on June 8, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

SERY AND HEHRE

JAFFE, E. 1941 Influence of sucrose upon

production of serologically reactive material

by certain streptococci. Proc. Soc. Exptl.

Biol. Med., 47, 339-344.NORDSTROM, L. AND HULTIN, E. 1948 Dex-

tranase, a new enzyme from mould. Svensk

Kem. Tidskr., 60, 283-284.

SUGG, J. Y. AND HEHRE, E. J. 1942 Reactionsof dextrans of Leuconostoc mesenteroides withthe antiserums of Leuconostoc and of types 2,20 and 12 pneumococcus. J. Immunol., 43,

119-128.

TSUCHIYA, H. M., BORUD, 0. J., AND CORMAN, J.1950 The synthetic action of an enzymatic

preparation of Aspergillus niger NRRL 330.

Abstracts of papers, 118th National A. C. S.

meeting.

TSUCHIYA, H. M., JEANES, A., BRICKER, H. M.,AND WILHAM, C. A. 1952a Dextran-degrad-

ing enzymes from molds. J. Bacteriol., 64,

513-519.

TSUCHIYA, H. M., KOEPSELL, H. J., CORMAN, J.,BRYANT, G., BOGARD, M. O., FEGER, V. H.,AND JACKSON, R. W. 1952b The effect of

certain cultural factors on production of

dextransucrase by Leuconostoc mesenteroides.J. Bacteriol., 64, 521-526.

WHITESIDE-CARLSON, V. AND CARLSON, W. W.

1952 Enzymatic hydrolysis of dextran.Science, 115, 43.

380 [VOL. 71

on June 8, 2018 by guesthttp://jb.asm

.org/D

ownloaded from